ViperEkura c0effc9f5b refactor: 位置编码改用 position_ids [B,S]，简化 attention mask 构建 ...

- RotaryEmbedding/CacheView 接受 position_ids 替代 start_pos

- process_attention_mask 用 position_ids >= arange 做逐位置 causal

- 训练/无 KV cache 时 position_ids=None 内部自动处理

- 移除 executor/benchmark 中冗余的 input_mask 构造

2026-05-14 13:26:31 +08:00

..

__init__.py

refactor: TaskManager 剥离页管理，STOP 移至 task.py

2026-05-11 14:04:31 +08:00

cache.py

refactor: 位置编码改用 position_ids [B,S]，简化 attention mask 构建

2026-05-14 13:26:31 +08:00

engine.py

chore: 解耦 Executor/Scheduler/TaskManager，修复 stop 页泄漏，移除 ServerState 全局单例

2026-05-12 13:47:55 +08:00

executor.py

refactor: 位置编码改用 position_ids [B,S]，简化 attention mask 构建

2026-05-14 13:26:31 +08:00

sample.py

refactor: TaskManager 剥离页管理，STOP 移至 task.py

2026-05-11 14:04:31 +08:00

scheduler.py

chore: 解耦 Executor/Scheduler/TaskManager，修复 stop 页泄漏，移除 ServerState 全局单例

2026-05-12 13:47:55 +08:00

server.py

chore: 解耦 Executor/Scheduler/TaskManager，修复 stop 页泄漏，移除 ServerState 全局单例

2026-05-12 13:47:55 +08:00

task.py

chore: 解耦 Executor/Scheduler/TaskManager，修复 stop 页泄漏，移除 ServerState 全局单例

2026-05-12 13:47:55 +08:00